I’m sharing my Synth V workflow here. I’m not sure if I’m on the right track.
#1. Switch to “Prefer Speed” for fast editing, then switch back to “Prefer Quality” for final rendering and export.
From my experience, “Prefer Speed” is about 3~4 times faster than “Prefer Quality”. However, I can not hear the difference between the two. The previews under two modes sound exactly the same to me. Why?
#2. Use AI auto pitch generation. No manual mode. No parameter editing.
Can I assume, everytime when I open the project and AI is regenerating the whole track, the result will be exactly the same? AI pitch generation should have no randomness each time, right?
In other words, if I make no global change, make no editing on a particular note, make no editing on the notes immediately before and after, then I can rest assure, that AI will not change the pitch curve of that note forever, am I right?
#3. Use AI Retake to fix problems such as vocal out of tune, over vibrato, etc.
I’m using Cong Zheng. Yes she is very passionate, I know, but she is always singing out of tune, and sounds very bad.
Instead of manually messing with the pitch curve and resulting in a robotic and unnatural sound, I choose to use AI Retake to fix these out of tune notes. Quickly generate 10 retakes, then simple select the best one, done.
However, I never generate retakes on multiple connected notes, but only doing it one note at a time. My understanding is, once you generate a retake on multiple notes, it is a combination, and you can not mix and match within a combo. Doing one note at a time will give you the flexibility to make your own combo.
Am I doing it right here?
#4. Everytime I edit a note, if it has AI retake on it, I will clear all retakes, and revert it back to default pitch.
My understanding is, everytime you make a change on a note, such as changing lyrics, position, length, etc, the AI will re-compute that note and give it a brand new pitch curve, which is unrelated to the previous one.
That’s why all previous generated AI retakes will change too. It’s safer to clear them all, start on a clean slate, generate them again, so they won’t associate with the previous note before the change was made.
Am I right here?
#5. AI retake will remember the pitch curve permanently.
Although generating a retake seems to be randomly done, but once that retake is generated, the pitch curve will be stored in the project file forever.
What I mean is, if you don’t make any changes on that note or adjacent notes, then each time you open the project and AI is reloading the waveform, it will keep the AI retake pitch curve unchanged as your previous session, right?
So, if I want to work on a project, and keep my notes 100% on auto mode, it’s totally safe, no pitch curve will be lost, right?
I never have to bother changing notes to manual mode for the purpose of “locking the existing pitch curve”, am I right?
Thanks to everyone for your thoughts for food.